503 Final Presentation
-
Upload
kklo -
Category
Technology
-
view
317 -
download
0
description
Transcript of 503 Final Presentation
![Page 1: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/1.jpg)
TIMELINE FROM NEWS
KK Lo
![Page 2: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/2.jpg)
GOAL...
![Page 3: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/3.jpg)
![Page 4: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/4.jpg)
RELATED WORK
![Page 5: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/5.jpg)
Topic Detection and Tracking
Temporal and Event Tagging
2communities
![Page 6: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/6.jpg)
Topic Detection and Tracking
tracking topics?classifying documents
discovering new topic
Events of interest
![Page 7: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/7.jpg)
assume each article is an event
Problems
lack of details
publication date =event happen time?
![Page 8: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/8.jpg)
Temporal and Event Tagging
? Tagging events and their temporal relationships
![Page 9: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/9.jpg)
too many Events....
Problems
Result obtained from the TARSQI toolkit
Event
Event Event
Event
EventEventEvent
![Page 10: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/10.jpg)
MY SOLUTION
![Page 11: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/11.jpg)
APPLY SUMMARIZATIONTECHNIQUE AS
EVENT FILTERING
![Page 12: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/12.jpg)
3components
![Page 13: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/13.jpg)
Prior Ranking1. Sentence A
2. Sentence B
3. Sentence C
4. ...
Beginning sentence has a higher prior probability
0prior probability
![Page 14: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/14.jpg)
Grasshopper
A Page-rank-like ranking algorithm
s1
s2s3
s4
s5
cosine similarities
![Page 15: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/15.jpg)
TARSQI Toolkit
explicit time
event instance
event-time link
event-event link
From TEXT to TimeML
![Page 16: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/16.jpg)
Event FilteringEvents in TimeML
Appear in the Top Selected Sentences?
PICK
BYENO
YES
![Page 17: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/17.jpg)
Temporal Reasoner
Find the (start, end) bound for each events
2008Dec
event1event2
event3
2009
![Page 18: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/18.jpg)
RESULT?
![Page 19: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/19.jpg)
Sentence Selection Quality
Special Thanks to
for the data and ROUGE =p
250-words summary form 25 documents with DUC2007 Data Set
![Page 20: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/20.jpg)
How can we represent 3320 events on a timeline?
Effect of Sentence Filtering
D0701A D0720E
#Event before Filtering 3320 1435
#Event after Filtering 67 37
choosing the top 10 sentences
![Page 21: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/21.jpg)
This shows that my approach is a failure
Time-Event AnchoringD0701A D0720E
#Event before Filtering
3320 1435
#Failure 3085 1129
#Event after Filtering
67 37
#Failure 49 29
![Page 22: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/22.jpg)
WHY?Unable to deduce the
relationships for all pair of events
TARSQI only support single document
e.g. 50 tagged events,only 50 pairs of relation are taggedshould be 50C2 = 1225
![Page 23: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/23.jpg)
LESSON LEARNED
![Page 24: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/24.jpg)
3areas
Topic Detection and Tracking
Temporal and Event Tagging
Automatic Summarization
my project
![Page 25: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/25.jpg)
The limit of existing technology
cannot get enough information from the documents
The limit of temporal analysis
OR EVEN
![Page 26: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/26.jpg)
cosine similarity with tf-idf weighting is computational
expensive
2.5 hrs for 867 sentences
![Page 27: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/27.jpg)
DUC2007 Documents are hard to parse
different documents have different format........
no standard date format...
contains some special characters that cause troubles
to XML parsers...
![Page 28: 503 Final Presentation](https://reader033.fdocuments.net/reader033/viewer/2022042817/559cf6281a28ab75438b4779/html5/thumbnails/28.jpg)
Q & A